Assessment of Density-Dependent Hydro-Collapse Mechanisms in Fine-Grained Geomaterials: A Multi-Axial Stress Analysis

Collapsing of soil is the sudden reduction in volume under a certain overburden stress upon wetting. Wetting-induced collapse settlement canCP = + ( ) − − + 47 506 0 072 0 4393 123 2 851 . . . . ln( )− wi d w.γ (2)where, CP is the collapse potential in percentage; Cu is the coefficient of uniformity; wi is the initial moisture content (%); γd is the initial dry unit weight (kN/m3); pw is the pressure at wetting (kPa); and (S − C) is the difference between sand and clay contents (%). However, Habibagahi and Taherian (2004) found the above equations gave poor performance in predicting collapse potential, especially over smaller collapse potentials. This might be due to the fact that the use of inappropriate parameters and overlooking of some important parameters like matric suction, which may have a significant role in maintaining the meta-stable loose structure of collapsible unsaturated soil.

Collapsible soils are a type of problematic soils which are found in many regions of Iran. These soils show sudden and irreversible reduction in volume with applied stress at constant moisture content or with increase in moisture content at a constant stress level or changes in both. The purpose of this research was to assess the geotechnical properties of loess soil around Mohammad-Abad of Semnan railway station and determine its collapsibility potential. Remolded soil samples were tested for their collapsibility using oedometer tests. In addition, the application of Cement Kiln Dust (CKD) to reduce collapsibility of two types of soils collected from this region was studied. The results indicate that the types of aggregation, clay content, salt, and porosity had a significant impact on the collapse phenomenon of the tested soils. The addition of CKD to both soils types resulted in more than 50% reduction in the collapsibility potential.

In this research, Artificial Neural Networks (ANNs) will be used in an attempt to predict collapse potential of gypseous soils. Two models are built one for collapse potential obtained by single oedemeter test and the other is for collapse potential obtained by double oedemeter test. A database of laboratory measurements for collapse potential is used. Six parameters are considered to have the most significant impact on the magnitude of collapse potential and are being used as an input to the models. These include the Gypsum content, Initial void ratio, Total unit weight, Initial water content, Dry unit weight, Soaking pressure. The output model will be the corresponding collapse potential. Multi-layer perceptron trainings using back propagation algorithm are used in this work. A number of issues in relation to ANN construction such as the effect of ANN geometry and internal parameters on the performance of ANN models are investigated. Information on the relative importance of the factors affecting the collapse potential are presented and practical equations for prediction of collapse potential from single oedemeter test and double oedemeter test in gypseous soils are developed. It was found that ANNs have the ability to predict the collapse potential from single oedemeter test and double oedemeter test in gypseous soil samples with a good degree of accuracy. The ANN models developed to study the impact of the internal network parameters on model performance indicate that ANN performance is sensitive to the number of hidden layer nodes, momentum terms, learning rate, and transfer functions. The sensitivity analysis indicated that for the models the results indicate that the initial void ratio and gypsum content have the most significant affect on the predicted the collapse potential.Keywords. Artificial Neural Networks, collapse potential, gypseous soils

In this research, Artificial Neural Networks (ANNs) is used in an attempt to predict collapse potential of gypseous soils. Two models are built; one for collapse potential obtained by single oedemeter test and the other is for collapse potential obtained by double oedemeter test. A database of laboratory measurements for collapse potential is used. Six parameters, which are 1.Gypsum content, 2.Initial void ratio, 3.Total unit weight, 4.Initial water content, 5.Dry unit weight, 6.Soaking pressure. are considered to have the most significant impact on the magnitude of collapse potential and used as an input to the models. The output model is the corresponding collapse potential. Multi-layer perceptron trainings using back propagation algorithm are used in this work. A number of issues in relation to ANN construction such as the effect of ANN geometry and internal parameters on the performance of ANN models are investigated. Information on the relative importance of the factors affecting the collapse potential are presented and practical equations for prediction of collapse potential of single oedemeter test and double oedemeter test in gypseous soils are developed. It is found that ANNs have the ability to predict the collapse potential of single oedemeter test and double oedemeter test in gypseous soil samples with a good degree of accuracy. The ANN models developed to study the impact of the internal network parameters on model performance indicate that ANN performance is sensitive to the number of hidden layer nodes, momentum terms, learning rate, and transfer functions. The sensitivity analysis indicated that the initial void ratio and gypsum content have the most significant affect on the prediction of collapse potential.

Collapsible soils are soils that are susceptible to a significant and sudden reduction in volume upon wetting and loading. This phenomenon is associated with loosely deposited, metastable soil structures, which can withstand a considerable stress in the unsaturated state, but they exhibit an excessive settlement when saturated. Collapsing of soil depends on the initial dry density, initial moisture content and pressure at wetting. Matric suction is also a key factor, which is related to moisture content through the soil water characteristic curve (SWCC). Suction helps to maintain the inter-granular contacts between particles of unsaturated soil. This study investigated the role of matric suction, dry density ratio and wetting pressure on the collapse behavior of three silty glacial sands from South Australia. The drying SWCC of each sand was determined by controlling matric suction between suctions of 1 to 1000 kPa, using the hanging column and pressure plate methods. The one-dimensional collapse potentials of the three different sands were determined at various specimen preparations and wetting pressures. It is shown that the collapse potential for a particular density and wetting pressure decreases generally with decrease in initial matric suction. An empirical equation to predict collapse potential of South Australian silty glacial sand was proposed using appropriate parameters. The equation was found to estimate collapse better than that of existing empirical equations.

This study investigated the relationship between pore characteristics and unit weight of clayey soils. This relationship was particularly important in case of expansive soils, as the pore characteristics determine moisture flux boundaries, which in turn represent volume change behavior. Current research tried to evaluate the effect of compaction unit weight on the pore size and pore volume along with consolidation and swell characteristics on two expansive clays from semi-arid environment. The two selected clays represent soils with different degrees of expansion, particle size and mineralogy. Pore size characterization for these two soils was performed using Mercury Intrusion Porosimetry, while swell and consolidation characteristics were determined using a conventional oedometer test. Samples for both tests were compacted at different unit weights including, 100, 95, 90, 80, 75, and 70 % of maximum dry unit weight (MDUW) obtained from standard proctor. The compaction water content was kept constant for all unit weight levels. Both pore volume and pore size distribution was analyzed with varying unit weight characteristics and particle sizes. In addition, swell strains and compression indices were studied with varying unit weight of compacted specimens. It was observed that, in the case of samples compacted at 100 % MDUW, about 50 % of the pores were larger than 0.1μm, and this value increased with reduction in unit weight. Current research is of practical importance, especially in the wake of microbial treatments for clayey soils where the passage of microbes depends on the pore size and more specifically pore throat size.

This paper analyzes the bearing capacity of two-layered soil medium using finite element (FE) software ABAQUS/CAE 2023. Although geotechnical engineers design foundations for layered soil, majorly current geotechnical studies emphasize single homogenous soil. So, this research has significant novelty as it focuses on layered soil and adds to the current literature. A nonlinear FE model was prepared and analyzed to determine the ultimate bearing capacity of two-layered soil (sandy soil over clayey soil). The Drucker–Prager and Mohr–Coulomb models were used to represent sandy soil and clayey soil layers, respectively. Strip footing material properties were considered isotropic and linearly elastic. This study performed parametric studies to understand the effects of thickness, unit weight, and the modulus of the elasticity of sandy soil on the ultimate soil bearing capacity. Additionally, it also analyzed the effect of the cohesive strength of clayey soil on layered soil bearing capacity. Results showed that an increase in sandy soil layer thickness strengthens the layered soil, and thus, improves the bearing capacity of soil. Increasing the sandy soil layer thickness over footing width (h1/B) ratio from 0.15 to 2.0 improved the ultimate bearing capacities with elastic settlements of 350 mm and 250 mm by 145.62% and 101.66%, respectively. Additionally, for a thicker sandy soil layer, an increase in the unit weight and modulus of the elasticity of sandy soil led to higher ultimate bearing capacity. Furthermore, it was concluded that an increase in clayey soil’s cohesive strength from 20 kPa to 30 kPa resulted in a 24.31% and 3.47% increase in soil bearing capacity for h1/B = 0.15 and h1/B = 2.0, respectively. So, the effect of cohesion is prevalent in the case of a thicker clayey soil layer.

Shrinkage is the volume reduction of a soil due to desiccation and decreasing pore pressure. This is important for the determination of volume based physical and hydraulic soil properties in the laboratory e.g., bulk density, volumetric moisture and water retention functions. Furthermore, it leads to changing surface elevation and crack formation at the field scale.There are two types of shrinking soils, clayey soils and organic soils, which are defined here as soils having a soil organic carbon content (SOC) above 7.5%. Clayey and organic soils differ strongly in their shrinkage behavior.  Furthermore, only few shrinkage studies differentiate between different organic soils. Parameters of existing shrinkage models are fully empiric and not directly linked to soil properties as dry bulk density, SOC, botanical composition and degree of decomposition.To determine the soil shrinkage characteristics (SSC i.e., relationship between moisture ratio and void ratio) of a variety of organic soils, we determined sample volumes with a three-dimensional (3D) structured light scanner at different moisture states from full saturation to dryness. We sampled 33 horizons (n = 4 replicates each) covering a wide range of botanical composition, development stages and degree of decompositions. Desiccation was performed by suction plates up to pressure heads of -200 hPa, followed by evaporation and oven-drying at 105°C. Volume and height of the 3D models created this way were determined by 3D graphic software and R, respectively. The volumetric moisture was determined by weighing the sample before and after scanning. Afterwards, volume and volumetric moisture were converted to moisture ratio and void ratio with the volume of solid particles. Due to small differences in particle volume between the replicates, both, moisture and void ratio were normalized by dividing them by the value at saturation. This normalization led to congruent results for the replicates.The shape of the SSCs strongly depended on the botanical composition and degree of decomposition. Peat consisting of slightly decomposed Sphagnum remains showed a supernormal shrinkage phase, i.e., volume reduction exceeds lost water volume at the dry end of the SSC and a relatively large range where volume reduction is (much) smaller than lost water volume, i.e., subnormal or structural shrinkage phase, at the wet end. The latter behavior was also shown by amorphous top soils. With increasing degree of decomposition or complete absence of Sphagnum remains the SSC flattened and tended to show a (near-) normal shrinkage phase, i.e., volume reduction equals lost water volume.The results showed that rigid Sphagnum remains strongly influence the shrinkage behavior of organic soils by stabilizing the matrix during desiccation until the large pores collapse rapidly when soil moisture and tension undercut a certain level. The SSC of organic soils without rigid fibers were more similar to SSCs of clayey soils.

The feasibility of using neural networks to model the complex relationship between soil parameters, loading conditions, and the collapse potential is investigated in this paper. A back propagation neural network process was used in this study. The neural network was trained using experimental data. The experimental program involved the assessment of the collapse potential using the one-dimensional oedometer apparatus. To cover the broadest possible scope of data, a total of eight types of soils were selected covering a wide range of gradation. Various conditions of water content, unit weights and applied pressures were imposed on the soils. For each placement condition, three samples were prepared and tested with the measured collapse potential values averaged to obtain a representative data point. This resulted in 414 collapse tests with 138 average test values, which were divided into two groups. Group I, consisting of 82 data points, was used to train the neural networks for a specific paradigm. Training was carried out until the mean sum squared error (MSSE) was minimized. The model consisting of eight hidden nodes and six variables was the most successful. These variables were: soil coefficient of uniformity, initial water content, compaction unit weight, applied pressure at wetting, percent sand and percent clay. Once the neural networks have been deemed fully trained its accuracy in predicting collapse potential was tested using group II of the experimental data. The model was further validated using information available in the literature. The data used in both the testing and validation phases were not included in the training phase. The results proved that neural networks are very efficient in assessing the complex behavior of collapsible soils using minimal processing of data.

Fill materials which have been inadequately compacted or placed excessively dry usually undergo a reduction in volume when their moisture content is increased. This phenomenon can occur without any increase in applied stress and is commonly termed collapse compression. The increase in moisture content can be caused either by downward infiltration of surface water or by a rising groundwater level, and the associated ground movements can have a serious effect on structures which have previously been built on the fill. Consequently, where building on a non-engineered fill is contemplated, the assessment of collapse potential should be one of the most important facets of the ground investigation and a primary objective of ground treatment should be largely, if not totally, to eliminate collapse potential. The specification and control of the placement and compaction of engineered fill should also aim to eliminate collapse potential. The identification and measurement of collapse potential are rarely easy as it is not feasible to obtain undisturbed samples of many waste fills and the commonly available in situ testing techniques do not always correlate well with collapse potential. Following an investigation of collapse compression by the Building Research Establishment, a methodology for identifying and measuring collapse potential in fills is proposed which includes a newly developed procedure for a borehole infiltration test. (A)

Many of the soils undergo volumetric changes due to the change in the water content. Swell-shrink and collapse behavior of those soils affects the stress state in soil and the interacted structures. Shrinkage in the soil produce cracks of different patterns, and affects the swelling potential in next wetting cycle. This study covers swelling and collapsing properties of four different soils from Mosul city. The changes in swelling and collapsing properties with respect to number of wetting and drying cycles have been investigated. Also, A shrinkage cracks have been studied with aid of digital image after each drying cycle. Number of segments and area of cracks calculated with aid of AutoCAD package. Results indicated that, the collapse potential is influenced by soil type (soil composition) and applied loads. As the applied loads increase the collapse potential increases. For sandy soil the collapse potential decreased with increasing wetting and drying cycles, and for the clayey soils, swell potential decreased while collapse potential increased with these cycles. It has been shown that the cracks increase with wetting-drying cycles. Larger values of percent crack area to the initial sample area has been observed in the soil that contain more clay content than other types of soils.

Residual soils are in the category of questionable soils which have been experienced in the arid and semi-arid climatic zones of the world. The conditions in these zones favour the development of most unsafe collapsible soils. At their dry natural state, they possess awesome stiffness and high apparent shear strength, however upon flooding, may demonstrate a remarkable reduction in volume, consequently deteriorate in strength and collapse. In this research, the collapse phenomenon of residual soil collected from three locations in Auchi, Northern Edo, Nigeria has been investigated on undisturbed specimens by utilizing single Oedometer test. The results obtained from Oedometer tests were utilized to form the database to develop the Artificial Neural Network model for the prediction of collapse potential induced by flood. The influences of flood, flooding pressure, void ratio, dry density and porosity on soil collapse have been investigated. Six input parameters (i.e. Flooding Pressure, Initial void ratio, Initial water content, Initial dry density, Liquid limit and Initial porosity) are considered to have the most noteworthy influences on the degree of collapse and have been utilized as the model’s inputs while the model output will be the equivalent collapse potential. The proposed network was developed using Microsoft Visual Studio 2010 and the MS.NET Framework 4.0 and source codes were written in C-Sharp (C#). A supervised learning was utilized to train the Back Propagation feed forward multi-layer ANN algorithm with the momentum coefficient and learning rate as its parameters. The prediction performance of the Artificial Neural Network model was assessed by utilizing the primary statistical criterion proposed by Shahin, et al., [1] such as the coefficient of correlation, R2, and the root mean square error, RMSE. The model outcomes demonstrated that it has the aptitude to predict the collapse potential from single Oedometer test in residual soil samples with a good degree of precision with coefficient of correlation, R2 = 0.856 and root mean square error, RMSE = 166.199.

Soil collapse is defined as a considerable reduction in soil volume when inundated under constantly applied pressure is known to be responsible for the failure of geotechnical structures such as highway/railway embankments and earth dams. Gene expression programming (GEP) is used as an artificial intelligence (AI) for the formulations of collapse potential in terms of the difference between the sand and clay percentages or the coefficient of uniformity, initial water content, initial dry unit weight, and wetting pressure in this paper. The experimental data available in the literature have been gathered for predicting collapse potential with the empirical formulations developed in the training and test sets of GEP-based models. Besides, additional experimental data derived from different literature are obtained to confirm the applicability and generalizability of the developed GEP-based formulations. The prediction performances of GEP models are compared to the experimental results and regression-based formulations proposed in the literature. These comparisons and statistical values obtained from analyses show that the GEP-based models are detected to be more effective methods to estimate the collapse potential. Moreover, a series of parametric analysis is conducted to perceive influences of input parameters on collapse potential by using GEP-based formulations.

Compressibility and collapse characteristics of arid saline sabkha soils

A soil water unit weight of 9.8 kN/m3 has been universally considered to quantify soil volumetric phase properties such as void ratio and degree of saturation, but has been found to greatly vary depending on soil type and the volume scale with which it is defined. Recent experimental and theoretical evidence has indicated that the unit weight of soil water can be significantly greater than 9.8 kN/m3 for clayey soils when gravimetric water content is less than 30%. A soil water unit weight as high as 18.8 kN/m3 is evident for some expansive soils at low water content. The significance of abnormally high water unit weight in quantifying soil phase volumes, saturation, and void ratio was experimentally assessed for various clayey soils and theoretically interpreted as a function of water content and soil type. For clayey soils with low liquid limit, average soil water unit weight can be as high as 12.5 kN/m3 for gravimetric water content less than 10%. This leads to an overestimation of liquid-phase saturation and void ratio by as much as 8% if a soil water unit weight of 9.8 kN/m3 is used. For clayey soil with a high liquid limit, the average soil water unit weight can be as high as 18.8 kN/m3 for water content less than 18%, leading to overestimation of liquid-phase saturation by as much as 36% and void ratio by as much as 20% if 9.8 kN/m3 is used. Charts were developed to estimate average soil water unit weight as a function of soil specific surface area and water content, and as a function of liquid limit and water content. The commonly used value of 9.8 kN/m3 for water unit weight can lead to significant errors in estimating phase volumes, void ratio, and saturation for clayey soils.